Straightforward synthesis of a derivative of purpurosamine C from D- galactose

Article abstract of DOI:10.1016/S0957-4166(98)00083-4

Purpurosamine C (1) is a component of the aminoglycoside antibiotic gentamicin C(1a). A derivative of 1 was synthesized from D-galactose via its 2-acetoxy-3,4,6-tri-O-acetyl glycal (3). Compound 3 undergoes glycosylation with 2-propanol in the presence of SnCl₄, with two successive allylic rearrangements of the double bond to give isopropyl 6-O-acetyl-3,4-dideoxy- α-D-glycero-hex-3-enopyranosid-2-ulose (7). Compound 7 was hydrogenated, and O-deacetylated to afford 8. The free OH group of 8 was tosylated and substituted by azide, and the carbonyl function of the resulting ulose 10 reacted with hydroxylamine to give the E,Z-oximes (11,12). Highly diastereoselective reduction of the oxime acetate (13) by borane, which also reduced the azide function, led to the purpurosamine C derivative 14 (~40% yield from 3).

Full text of DOI:10.1016/S0957-4166(98)00083-4

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 1269–1276
Straightforward synthesis of a derivative of purpurosamine C
from D-galactose
Patricia A. Zunszain and Oscar Varela ∗
CIHIDECAR (CONICET)— Departamento de Química Orgánica, Facultad de Ciencias Exactas y Naturales, Universidad de
Buenos Aires, Ciudad Universitaria, Pabellón 2, 1428 Buenos Aires, Argentina
Received 13 January 1998; accepted 15 February 1998
Abstract
Purpurosamine C (1) is a component of the aminoglycoside antibiotic gentamicin C1a. A derivative of 1 was
synthesized from D-galactose via its 2-acetoxy-3,4,6-tri-O-acetyl glycal (3). Compound 3 undergoes glycosylation
with 2-propanol in the presence of SnCl4, with two succesive allylic rearrangements of the double bond to give
isopropyl 6-O-acetyl-3,4-dideoxy-α-D-glycero-hex-3-enopyranosid-2-ulose (7). Compound 7 was hydrogenated,
and O-deacetylated to afford 8. The free OH group of 8 was tosylated and substituted by azide, and the
carbonyl function of the resulting ulose 10 reacted with hydroxylamine to give the E,Z-oximes (11,12). Highly
diastereoselective reduction of the oxime acetate (13) by borane, which also reduced the azide function, led to the
purpurosamine C derivative 14 (∼40% yield from 3). © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
Purpurosamine C (1) is one of the components of the broad spectrum antibiotic gentamicin C1a,
produced by Micromonospora purpurea.1 It is also a constituent of semisynthetic aminoglycoside
antibiotics of the 30,40-dideoxykanamycin type, such as dibekacin and arbekacin.2,3 Several syntheses
of derivatives of this diaminotetradeoxyhexose have been reported. The sigmatropic rearrangement
of an allylic thiocyanate4 (which isomerizes to isothiocyanate) or azide5–7 was the key step in some
syntheses. Levoglucosenone8 and glucosamine9,10 have been employed as starting materials for the
synthesis of derivatives of purpurosamine C. Also, non-carbohydrate precursors have been used to
prepare purpurosamine C in racemic form. Thus, addition of nitrosyl chloride to 2-acetoxymethyl-3,4-
dihydro-2H-pyran,11 or [4+2] cycloaddition of a diene to a 2-amino aldehyde12 led to the diaminosugar.
Alternatively, addition of N-glyoxyloyl-(2R)-bornane-10,2-sultam to a diene,13 or cycloaddition from a
carbohydrate precursor,14 led to enantiomerically pure purpurosamine C derivatives.
∗
Corresponding author. E-mail: varela@qo.fcen.uba.ar
0957-4166/98/$19.00 © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00083-4

1270
P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
The above-mentioned syntheses are commonly multi-step procedures and the overall yields are
relatively low. We wish to report here a straightforward route to purpurosamine C, starting from
inexpensive and commercially available galactose, via its 2-acetoxy glycal triacetate (3).
2. Results and discussion
The retrosynthetic analysis of purpurosamine C (Scheme 1) suggests a dideoxyulose derivative
(2) as a convenient synthon, since the 6-OH group can be readily converted into an amine and the
carbonyl at C-2 is a good precursor of the other amino group. The synthesis of such an ulose from a
monosaccharide would require a rather complex sequence of selective protections of hydroxyl groups,
oxidation, elimination and reduction. However, we have found in our laboratory15 that Lewis acid
catalyzed glycosylation of 2-acyloxyglycals leads to enulose derivatives in a single step. Thus, 2-
acetoxy-3,4,6-tri-O-acetyl-D-galactal (3), readily obtained from D-galactose,16 underwent the tin(IV)
chloride promoted glycosylation to give the enulose 4, in 90% yield. Compound 4 arises from the
addition of 2-propanol to the anomeric center with Ferrier’s rearrangement of the double bond.17 The
vinylic acetate of the resulting enopyranoside is attacked by the alcohol promoting a second allylic
rearrangement, with elimination of the acetoxy substituent at C-4 and generation of the enone system.
The remarkable diastereoselectivity in the formation of the glycosylic linkage may be attributed to
the anchimeric participation of the acetoxy group at C-2. This group can stabilize the positive charge
at C-1 by formation of an acyloxonium ion having a β-configuration, because of the stereoelectronic
requirement of a positively charged anomeric substituent18 (reverse anomeric effect). Therefore, the
attack of the alcohol should take place from the opposite face, leading to the α-glycosylic bond.
Scheme 1.
Attempted deacetylation of 4 afforded the desired derivative 5, having the free hydroxyl group at C-6,
together with a considerable amount of the Michael addition product 6 (Scheme 2). The configuration of
1
the stereocenter at C-4 of the major isomer was determined to be R by means of its H NMR spectrum
0
(Table 1). Thus, the small J3,4 and J3 ,4 (3.4 and 3.1 Hz, respectively) indicated that H-4 bisects the
H–C(3)–H0 angle, and hence that the methoxy substituent is quasiaxially oriented. This result is also
confirmed by the small value of J4,5 (∼1 Hz). In order to avoid the formation of 6, compound 4 was
hydrogenated to give 7, which was treated with sodium methoxide to afford 8 (93% yield from 4). The
13C NMR spectrum of 8 (Table 2) showed the resonances of the anomeric and the carbonyl carbons at
202.9 and 97.8 ppm, respectively. The signals of the carbons bonded to oxygen (C-5 and C-6) appeared at
68.8 and 64.8 ppm, and those of the methylene carbons at 34.7 and 28.7 ppm. Compound 8 is identical to
2, the direct precursor of purpurosamine proposed in the retrosynthetic analysis (Scheme 1). Tosylation
of crude 8 gave 9 in 94% overall yield from 7 (Scheme 3).
Nucleophilic displacement of the tosylate group of 9 by sodium azide in DMF at 80°C (18 h) led to the
azide derivative 10 in 82% yield. The substitution of tosylate by azide produces in the 13C NMR spectra
a strong shift for the C-6 signal from 71.0 to 54.2 ppm. The 1H NMR spectra of 9 and 10 also showed
significant changes. For example, the double doublets of the methylene protons vicinal to the tosylate
(4.09 and 4.02 ppm) appeared shifted upfield in the azide (3.35 and 3.25 ppm).

P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
1271
Scheme 2.
Table 1
1 H NMR data for compounds 5–14
The reaction of 10 with hydroxylamine hydrochloride in the presence of sodium hydrogen carbonate
led to two products which had different mobility by TLC (Rf 0.45 and 0.31; toluene:EtOAc=9:1). This
fact indicated that, as expected for non-symmetrical ketones, a diastereoisomeric (syn, anti) mixture
of ketoximes had been obtained. The mixture was separated by column chromatography, and each
1
compound was individually characterized. The H and 13C NMR spectra of both products showed a
pattern consistent with that of an oxime structure. The absolute configuration was established taking into
account the shift in the resonances of the carbons α to the imine group. These shifts are associated with
the conversion of a ketone into its oxime, with the effect for the α-syn carbon being greater than for the
α-anti carbon.19 With the aid of this correlation, it was easy to assign as the E (anti)-oxime (11), the
compound which eluted first from the column. Thus, the signals of C-1 and C-3, when compared with

1272
P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
Table 2
13C NMR data for compounds 5–14
Scheme 3.
the resonances of the same carbons in 10, showed a shifting of 2.4 and 16.3 ppm, indicating that C-3 is
syn to the N–OH group, and C-1 is anti. Consistent with this assignment, the other oxime (12) had the
Z-configuration, according to the same rule.
On standing in a chloroform solution, the Z-isomer (12) progressively converts into the E-isomer (11),
probably due to the lower steric hindrance of the OH group syn to the least substituted α-carbon (C-3).
For ketoximes where two isomers may be expected, the major isomer is that of the anti configuration.19
Various reagents and reaction conditions were employed for the reduction of oximes to amines.
Hydrogenolysis of the mixture of 11 and 12 in the presence of Pd or Pt catalysts, with different solvents,
concentrations and pressures, gave extremely poor yields of the diamine. Also, reduction with lithium
aluminun hydride or Raney nickel proved to be unsuccesful. Borane has been described as a good
reducing agent for the preparation of amines from oximes20 at high temperatures. However, the reduction
of oxime ethers or oxime esters can be accomplished at room temperature.20,21 In the sugar field,
Danishefsky and Maring22 performed the borane reduction of a ketoxime acetate into amine. Therefore,

P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
1273
the mixture of oximes 11 and 12 was acetylated under standard conditions and the acetyl derivative 13
was treated with an excess of borane in THF. We expected that the azide function would be also reduced to
amine under those conditions.23 According to our expectations, after acetylation of the reduction product,
the N,N0-diacetylpurpurosamine derivative 14 was obtained. The 1H NMR spectrum of 14 showed the
signals of two NH protons at 5.85 and 5.56 ppm and the singlets of two CH3 acetyl groups, which dictates
that both the azido and the oxime have been reduced. The stereochemistry at C-2 was established on the
basis of J1,2, J2,3eq and J2,3ax (3.4, 4.0 and 12.1 Hz, respectively) that clearly showed that the amino
group was equatorially oriented. The remarkable diastereoselectivity observed for the saturation of the
imine double bond may be attributed to the steric hindrance of the quasiaxially disposed vicinal anomeric
isopropyl group, which blocks the α-face and induces the attack of borane from the opposite side. We
have previously observed similar anomeric stereocontrol in the reduction of a vicinal carbonyl group.24
In summary, a simple and direct synthesis of purpurosamine C derivative 14 has been accomplished
using readily available reagents, in an overall yield of ∼40% (from 2-acetoxy-3,4,6-tri-O-acetyl-D-
galactal). Some key features of the synthesis are:
(i) the chiral template for the synthesis of 14, the dideoxyhex-2-ulose 7, was readily prepared from
per-O-acetyl-D-galactose by a high yielding, three-step procedure;
(ii) the nitrogen containing groups (azide and oxime) were easily introduced in the molecule of 7 by
direct functional group transformations and with excellent yields; and
(iii) the reduction of the azide and the oxime was performed in a single, highly diastereoselective step,
to afford the purpurosamine C derivative 14.
3. Experimental
3.1. General
Solvents were dried and purified by appropriate standard procedures. TLC was performed on aluminun
plates precoated with silica gel 60 F254 (Merck), developed with the solvents indicated in each individual
case. The spots were visualized by exposure to UV light and by heating after immersion in a solution
of p-anisaldehyde (0.5% v/v) in ethanol containing 5% H2SO4. Column chromatography was performed
with silica gel 60 (200–400 mesh, Merck). Optical rotations were measured with a Perkin–Elmer 343
polarimeter. Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker AC 200 (1H: 200
MHz, 13C: 50.3 MHz) or a Bruker 500 (1H: 500 MHz) spectrometer in CDCl3 with TMS as an internal
standard unless otherwise stated. Melting points are uncorrected and were determined on a Fisher–Johns
apparatus. Mass spectral data, electron impact mass spectra (EIMS) and FAB+, were obtained on a
Shimadzu QP5000 spectrometer opperating at 70 eV, and a ZAB-SEQ4F spectrometer, respectively.
3.2. 2-Acetoxy-3,4,6-tri-O-acetyl-D-galactal (3) and isopropyl 6-O-acetyl-3,4-dideoxy-α-D-glycero-
hex-3-enopyranosid-2-ulose (4)
Compound 3 was prepared from per-O-acetyl-D-galactose as previously described.16 Treatment of 3
with 2-propanol in the presence of tin(IV) chloride afforded 4 (90% yield).15

1274
P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
3.3. Isopropyl 3,4-dideoxy-α-D-glycero-hex-3-enopyranosid-2-ulose (5) and isopropyl 3-deoxy-4-O-
methyl-α-D-threo-hexopyranosid-2-ulose (6)
To an externally cooled (0°C) and stirred solution of 4 (121 mg, 0.54 mmol) in anhydrous methanol
(30 mL) was added 0.5 M NaOMe (1 mL) under N2. After 1 h of stirring, TLC showed complete
transformation of the starting material into two lower moving compounds having Rf 0.29 and 0.19
(toluene:EtOAc=2:1). The solution was neutralized by addition of Dowex 50W (H+). The resin was
filtered and the filtrate was concentrated to a syrup, which was chromatographed on a silica gel column
with toluene:EtOAc (5:1) as eluent. The 2-ulose 5 (31 mg, 31%) was isolated from the fractions of Rf
0.29. EIMS, m/z: 155 (M+·−CH2OH·, 1%), 127 (M+·−OiPr·, 3%), 114 (3%), 113 (4%), 98 (54%), 97
(38%), 85 (9%), 81 (9%), 70 (41%), 55 (20%), 42 (100%). FAB, m/z: 187 (M+1).
The more polar product was identified as a 9:1 mixture of 6 and its erythro isomer (48 mg, 42%).
EIMS, m/z: 190 (M+·−CO·, 1%), 175 (M+·−iPr·, 1%), 159 (M+·−OiPr·, 11%), 130 (27%), 115 (7%),
99 (7%), 97 (7%), 89 (8%), 87 (12%), 84 (11%), 71 (100%), 61 (37%), 58 (33%), 57 (23%), 55 (11%),
45 (24%). FAB, m/z: 219 (M+1).
3.4. Isopropyl 3,4-dideoxy-α-D-glycero-hexopyranosid-2-ulose (8)
To a stirred solution of 724 (0.54 g, 2.3 mmol) in anhydrous methanol (21 mL) was added 0.5
M NaOMe (2 mL) in an N2 atmosphere. After 3 h of stirring at room temperature the solution was
neutralized by addition of Dowex 50W (H+). The resin was filtered and the solvent evaporated to give a
syrup, which was chromatographed using toluene:EtOAc (3:1) to afford pure compound 8 (0.43 g, 97%);
[α]D=+128 (c 0.87, CHCl3). Anal. calcd for C9H16O4: C, 57.41; H, 8.57. Found: C, 57.30; H, 8.74.
3.5. Isopropyl 6-O-(p-tolylsulfonyl)-3,4-dideoxy-α-D-glycero-hexopyranosid-2-ulose (9)
To a stirred solution of crude 8 (0.45 g, 2.37 mmol) in dry CHCl3 (4.2 mL), externally cooled at 0°C,
were added pyridine (0.38 mL, 4.7 mmol) and tosyl chloride (0.7 g, 3.5 mmol). After stirring at room
temperature for 16 h, water (6 mL) was added, and the stirring was maintained for 0.5 h. The mixture was
diluted with CH2Cl2 and it was successively washed with 2 N aqueous HCl, saturated aqueous NaHCO3,
and water. The organic extract was dried (MgSO4) and concentrated. The resulting syrup was purified
by column chromatography with CH2Cl2 as eluent, to afford pure 9 (0.63 g, 94%); [α]D=+63 (c 0.98,
CHCl3). Anal. calcd for C16H22O6S: C, 56.12; H, 6.48. Found: C, 56.35; H, 6.62.
3.6. Isopropyl 6-azido-3,4,6-trideoxy-α-D-glycero-hexopyranosid-2-ulose (10)
Compound 9 (0.28 g, 0.82 mmol) was dissolved in anhydrous DMF (5 mL) and sodium azide (0.15
g, 2.3 mmol) was added. The resulting suspension was heated with stirring at 80°C for 18 h under N2,
at which time TLC showed a main spot having Rf 0.38 (hexane:EtOAc=10:1) and no starting material
remaining. The reaction mixture was diluted with CH2Cl2 and washed with water. The organic extract
was dried (MgSO4) and concentrated to a syrup, which was chromatographed using toluene:EtOAc (14:1)
as eluent. Compound 10 (0.14 g, 82%) was obtained as an oil; [α]D=+74 (c 0.81, CHCl3). Anal. calcd
for C9H15O3N3: C, 50.68; H, 7.09; N, 19.71. Found: C, 50.67; H, 6.76; N, 19.86.

P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
1275
3.7. Isopropyl 6-azido-3,4,6-trideoxy-α-D-glycero-hexopyranosid-2-ulose E- and Z-oximes (11 and 12)
To a solution of 10 (96 mg, 0.45 mmol) in anhydrous isopropyl alcohol (2 mL), hydroxylamine
hydrochloride (68 mg, 0.9 mmol) and NaHCO3 (72 mg, 0.87 mmol) were added, and the mixture was
stirred at room temperature for 1 h. TLC showed the complete transformation of the starting material
into two lower moving compounds having Rf 0.45 and 0.31 (toluene:EtOAc=9:1). The salts were filtered
and the solvent was evaporated to a syrup, which was chromatographed on a silica gel column with
toluene:EtOAc (10:1). The less polar product was characterized as the oxime having the E-configuration
(11, 56 mg, 55%); [α]D=+106 (c 0.70, CHCl3). Anal. calcd for C9H16O3N4: C, 47.34; H, 7.07; N, 24.55.
Found: C, 47.55; H, 7.21; N, 24.21. From the fractions of Rf 0.31 was isolated the Z-oxime (12, 40 mg,
39%), which converts into the E-isomer on standing.
3.8. Isopropyl 6-azido-3,4,6-trideoxy-α-D-glycero-hexopyranosid-2-ulose E,Z-oxime acetate (13)
The crude mixture of oximes obtained as described above from 10 (84 mg, 0.39 mmol) was dissolved
in pyridine (1 mL) and cooled to 0°C. Acetic anhydride was added (1 mL) and the mixture was stirred
for 1 h. The solvents were evaporated by co-distillation with MeOH and toluene, and the residue was
chromatographed (toluene:EtOAc=15:1) to give a ∼7:1 mixture of the E/Z oxime acetates (12, 102 mg,
96%). EIMS, m/z: 211 (M+·−OiPr·, 2%), 169 (25%), 113 (9%), 95 (9%), 84 (5%), 78 (4%), 68 (26%),
55 (16%), 43 (100%). FAB, m/z: 271 (M+1).
3.9. Isopropyl 2,6-N,N-diacetamido-2,3,4,6-tetradeoxy-α-D-erythro-hexopyranoside (14)
To a solution of compound 13 (101 mg, 0.37 mmol) in anhydrous THF (1.5 mL) cooled at −18°C, 2.5
M solution of BH3·THF (1.2 mL, 3 mmol) was added. The mixture was stirred for 3 h at −18°C, and
then at room temperature for an additional 18 h. Methanol (3 mL) was added to decompose the excess
of borane, and the solvents were evaporated. The residue was dissolved in trifluoroacetic acid (0.9 mL)
and the solution was stirred for 0.5 h at room temperature. TFA was evaporated, and the residue was
dissolved in pyridine (1 mL) and cooled to 0°C. Acetic anhydride (1 mL) was added, and the reaction
mixture was stirred at room temperature for 20 h. After evaporation of the solvents, the resulting syrup
was chromatographed using EtOAc to afford crystalline 14 (66 mg, 65%). After recrystallization from
acetone, 14 had m.p. 195–196°C; [α]D=+113 (c 0.71, CHCl3). Anal. calcd for C13H24O4N2: C, 57.37;
H, 8.88; N, 10.29. Found: C, 57.29; H, 9.04; N, 10.12.
Acknowledgements
We are indebted to UMYMFOR (FCEN-CONICET) for the microanalyses. Financial support from
the Universidad de Buenos Aires (Project EX 173), Fundación Antorchas and the National Research
Council of República Argentina (CONICET) is gratefully acknowledged. O.V. is a Research Member of
CONICET.
References
1. Umezawa, S. Adv. Carbohydr. Chem. Biochem. 1974, 30, 111.
2. Kuwahara, R.; Tsuchiya, T. Carbohydr. Res. 1997, 299, 271.

1276
P. A. Zunszain, O. Varela / Tetrahedron: Asymmetry 9 (1998) 1269–1276
3. Kondo, S.; Iinuma, K.; Yamamoto, H.; Maeda, K.; Umezawa, H. J. Antibiot. 1973, 26, 412.
4. (a) Guthrie, R. D.; Williams, G. J. Chem. Commun. 1971, 923. (b) Guthrie, R. D.; Williams, G. J. J. Chem. Soc., Perkin
Trans. 1 1972, 2619.
5. (a) Cleophax, J.; Leboul, J.; Olesker, A.; Gero, S. D. Tetrahedron Lett. 1973, 4911. (b) Cleophax, J.; Gero, S. D.; Jegou-
Aumont, E.; Leboul, J.; Mercier, D. J. Chem. Soc., Chem. Commun. 1975, 11.
6. Brimacombe, J. S.; Da’aboul, I.; Tucker, L. C. N. J. Chem. Soc., Perkin Trans. 1 1975, 979.
7. Cleophax, J.; Olesker, A.; Rolland, A.; Gero, S. D.; Forchioni, A. Tetrahedron 1977, 33, 1303.
8. Brimacombe, J. S.; Hunedy, F.; Mather, A. M.; Tucker, L. C. N. Carbohydr. Res. 1979, 68, 231.
9. Ohashi, Y.; Okuno, S.; Takeda, K.; Ito, Y. Carbohydr. Res. 1978, 67, 503.
10. Liang, X.; Krieger, R.; Prinzbach, H. Tetrahedron Lett. 1995, 36, 6433.
11. Brimacombe, J. S.; Da’aboul, I.; Tucker, L. C. N. J. Chem. Soc., Perkin Trans. 1 1974, 263.
12. Golebiowski, A.; Jacobsson, U.; Chmielewski, M.; Jurczak, J. Tetrahedron 1987, 43, 599.
13. Bauer, T.; Jezewski, A.; Jurczak, J. Tetrahedron: Asymmetry 1996, 7, 1405.
14. David, S.; Lubineau, A.; Gero, S. D. J. Org. Chem. 1979, 44, 4986.
15. Varela, O.; De Fina, G. M.; de Lederkremer, R. M. Synthesis 1988, 891.
16. Varela, O.; De Fina, G. M.; de Lederkremer, R. M. Carbohydr. Res. 1987, 167, 187.
17. Ferrier, R. J. Adv. Carbohydr. Chem. Biochem. 1969, 24, 199.
18. Perlin, C. L. Tetrahedron 1995, 51, 11901.
19. Hawkes, G. E.; Herwig, K.; Roberts, J. D. J. Org. Chem. 1974, 39, 1017.
20. Feuer, H.; Braunstein, D. M. J. Org. Chem. 1969, 34, 1817.
21. Hassner, A.; Catsoulacos, P. Chem. Commun. 1967, 590.
22. Danishefsky, S. J.; Maring, C. H. J. Am. Chem. Soc. 1985, 107, 1269.
23. Hassner, A.; Levy, L. A. J. Am. Chem. Soc. 1965, 87, 4203.
24. De Fina, G.; Varela, O.; de Lederkremer, R. M. Carbohydr. Res. 1993, 246, 1993.